By: Howard Chu (hyc.delete@this.symas.com), June 30, 2007 5:09 am
Room: Moderated Discussions
Dean Kent (dkent@realworldtech.com ) on 6/29/07 wrote:
---------------------------
>jschlossel (js@schlossel.com) on 6/29/07 wrote:
>---------------------------
>>I cancelled my Dell order and Im really glad we dont Core2 processors running management
>>systems in planes, trains, and automobiles.
>
>You are absolutely correct. It is much better to have processors with unknown
>bugs in these applications, as 'feeling safe' is much better than actually being safe. (grin).
Heh... Well, my car ('97 Probe GTS) runs on a Motorola 68HCxx family chip. I downloaded all the docs including errata with no hassle, and as I recall there really wasn't much to speak of anyway, those things just work. Likewise the code for car engine computers doesn't change much over time; the same algorithms get used for several model years, just with different parameter tables tuned to each engine. (I also had a 2 year subscription to Ford's EEPROM updates along with their NG STAR diagnostic/programming tool. Cancelled it after a year because there really weren't many meaningful differences in the updates.) They still had the occasional software bug here and there but overall that stuff was solid. The processors just did their work. The software bugs tended to show up (as expected) when new systems were integrated into a controller. Things like reading sensors, metering air/fuel and controlling spark timing were all mature chunks of code that had been written/tested/left alone. (The Probe, of course, had a Mazda engine, but you could trace its ignition code lineage back to a version licensed from Bosch back in the 80s...)
---------------------------
>jschlossel (js@schlossel.com) on 6/29/07 wrote:
>---------------------------
>>I cancelled my Dell order and Im really glad we dont Core2 processors running management
>>systems in planes, trains, and automobiles.
>
>You are absolutely correct. It is much better to have processors with unknown
>bugs in these applications, as 'feeling safe' is much better than actually being safe. (grin).
Heh... Well, my car ('97 Probe GTS) runs on a Motorola 68HCxx family chip. I downloaded all the docs including errata with no hassle, and as I recall there really wasn't much to speak of anyway, those things just work. Likewise the code for car engine computers doesn't change much over time; the same algorithms get used for several model years, just with different parameter tables tuned to each engine. (I also had a 2 year subscription to Ford's EEPROM updates along with their NG STAR diagnostic/programming tool. Cancelled it after a year because there really weren't many meaningful differences in the updates.) They still had the occasional software bug here and there but overall that stuff was solid. The processors just did their work. The software bugs tended to show up (as expected) when new systems were integrated into a controller. Things like reading sensors, metering air/fuel and controlling spark timing were all mature chunks of code that had been written/tested/left alone. (The Probe, of course, had a Mazda engine, but you could trace its ignition code lineage back to a version licensed from Bosch back in the 80s...)